RE: FW: [ws-rx] PR Issue 22: concrete proposal

["Gilbert Pilz" <gpilz@bea.com>]

Doug,

 

Forgive me if I am not understanding you correctly, but are you saying that it is a requirement that the AS and AD must be unaffected by their use of WS-RM? If so, this is the first time I have heard such a requirement put forth. It seems like a rather strange requirement. I've always imagined that, if I were writing (or re-factoring) an application to use WS-RM, there would be a number of things I would do differently. For example, I probably wouldn't bother with any sort of retry/resend logic on the client-side.

 

There could be any number of reasons that an RMD might wish to know that it has not received all of the messages in a Sequence. It might wish to log it, raise an alert, attempt some diagnostic procedures, or (*gasp*) inform the AD. Some rules:

 

1.) WS-RM cannot specify how or even *if* the RMD should inform the AD that it detected an incomplete sequence.

 

2.) A WS-RM implementation that chooses not to communicate the completeness/incompleteness of a sequence to the AD should still be considered as conforming to WS-RM (all other things being equal).

 

3.) If I write an application that depends upon a particular implementation of WS-RM to inform me that some of the messages sent by the client didn't make it through, I must accept the fact that my application may not port well to WS-RM implementations that don't provide this functionality.

 

The main point is that if you don't consider LastMsgNumber to be useful, you are free to implement an RMD that *ignores* it. We happen to think it is a useful idea, and we don't think requiring the RMS to add this information to TerminateSequence (and/or CloseSequence) places an undo burden upon anybody.

 

- gp

---

From: Doug Davis [mailto:dug@us.ibm.com]
Sent: Tuesday, October 31, 2006 11:28 AM
To: Gilbert Pilz
Cc: ws-rx@lists.oasis-open.org
Subject: Re: FW: [ws-rx] PR Issue 22: concrete proposal

RMS sent 5 messages, it knows whether or not it got back all 5 acks - it doesn't need an additional "did you get _all_ of my messages" marker when Ack1-5 says the same thing.  From the RMD perspective I still claim there will have to be some application logic/flags that says how many messages are in the "set of messages" that is totally indepdent of RM - because otherwise how did the AD know that there are only 5 messages in the case when RM is not being used at all?  The argument that the application requires this data implies that the scenario will only work when RM is turned on - and I don't buy that.  W.r.t the RMD needing this info - as I said, I think the only case _it_ needs it is when IncompleteSeqBehavior is set to "discardEntireSequence".  All other cases it just becomes interesting (but useless) info.

thanks,
-Doug

"Gilbert Pilz" <gpilz@bea.com>

10/31/2006 02:09 PM

To	<ws-rx@lists.oasis-open.org>
cc
Subject	FW: [ws-rx] PR Issue 22: concrete proposal

Forwarded from Hal . .

---

From: Hal Lockhart
Sent: Tuesday, October 31, 2006 8:48 AM
To: Gilbert Pilz
Subject: RE: [ws-rx] PR Issue 22: concrete proposal
 
I think that it is an absolute requirement that a reliable message delivery service provide a mechanism by which both the sender and receiver can determine if all the data has been received or not. Period.
 
I think this is such an obvious, intuitive requirement, that anyone not involved in this debate would immediately agree.
 
{Sir, I am starting a new company called Reliable Package Delivery. For a premium fee, we will deliver packages to you completely undamaged, but we refuse to tell you whether or not you received all the packages we were given which were addressed to you. Can I sign you up as a customer?}
 
The only reason for invoking IncompeteSequenceBehavior, was because an argument had been put forward that somehow knowing if the data was complete or not was “Application Information” and thus out of scope of the spec. The existence of IncompeteSequenceBehavior demonstrates conclusively that the TC does not consider this out of scope. It was never intended as an argument that the only reason you need to know whether or not you got all the data is to decide if you need to invoke the IncompeteSequenceBehavior.
 
Hal
 
 

---

From: Doug Davis [mailto:dug@us.ibm.com]
Sent: Monday, October 30, 2006 4:28 PM
To: Gilbert Pilz
Cc: ws-rx@lists.oasis-open.org
Subject: RE: [ws-rx] PR Issue 22: concrete proposal

Well, I think there may need to be text that says.... if IncompleteSequenceBehavior is 'discardEntireSequence' then either Close or Terminate is required - otherwise, as you've noted, the RMD will never be able to send on any messages to the AD.  Any other value of IncompletSeqBehavior doesn't have this requirement though.
thanks,
-Doug

"Gilbert Pilz" <gpilz@bea.com>

10/30/2006 06:55 PM

To	Doug Davis/Raleigh/IBM@IBMUS
cc	<ws-rx@lists.oasis-open.org>
Subject	RE: [ws-rx] PR Issue 22: concrete proposal

 

Doug,
 
W/respect to LastMsgNumber in CloseSequence; I can see your point. I wouldn't object if you ammended our proposal to include LasMsgNumber in CloseSequence.
 
W/respect to TerminateSequence being required; if you ammend our proposal to optionally include LasMsgNumber in CloseSequence, then the requirement to always send TerminateSequence would go away. Clearly if the CS message contains LastMsgNumber then the RMD will have enough information to determine if it received all the messages in the sequence.
 
- gp

---

From: Doug Davis [mailto:dug@us.ibm.com]
Sent: Monday, October 30, 2006 10:10 AM
To: Gilbert Pilz
Cc: ws-rx@lists.oasis-open.org
Subject: Re: [ws-rx] PR Issue 22: concrete proposal


Gil,
If we head this route I'd like to see a LastMsgNumber on the Close as well so that we don't need to wait for a TerminateSequence before we can know the final outcome of the Sequence - w.r.t. what we can deliver to the app (ie. incompleteSeqBehvaior).  This would mean that either Close or Terminate would be required based on certain IncompletSeqBehavior values.
Also, on TerminateSequence being required....what does that mean w.r.t. enforcement?  Can we enforce it?  What happens if it not sent? Seems like its only really required for "DiscardEntireSequence" values of IncompleteSeqBehavior, all others can still do their job w/o it, no?  Would it not make more sense to say that its required only when needed by this flag to do its job?
thanks
-Doug

"Gilbert Pilz" <gpilz@bea.com>

10/26/2006 03:48 PM

 

To	<ws-rx@lists.oasis-open.org>
cc
Subject	[ws-rx] PR Issue 22: concrete proposal

 

Attached is a proposal for PR i022 in the form of a diff against CD-04.
The main points are:

1.) wsrm:TerminateSequence has been expanded to include a mandatory
LastMsgNumber element the value of which is, surprisingly enough, the
number of the last message in the Sequence.

2.) Sending wsrm:TerminateSequence is now mandatory; basically the whole
thing won't hold together unless the RMS is required to send a
wsrm:TerminateSequence.

<<wsrm-1.1-spec-pr-i022.pdf>>
[attachment "wsrm-1.1-spec-pr-i022.pdf" deleted by Doug Davis/Raleigh/IBM]